submission_id: neversleep-llama-3-lumim_8666_v2
developer_uid: zonemercy
status: inactive
model_repo: neversleep/llama-3-lumimaid-8b
reward_repo: ChaiML/reward_gpt2_medium_preference_24m_e2
generation_params: {'temperature': 1.0, 'top_p': 1.0, 'min_p': 0.0, 'top_k': 40, 'presence_penalty': 0.0, 'frequency_penalty': 0.0, 'stopping_words': ['\n'], 'max_input_tokens': 512, 'best_of': 4, 'max_output_tokens': 64}
formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}:', 'truncate_by_message': False}
reward_formatter: {'memory_template': "{bot_name}'s Persona: {memory}\n####\n", 'prompt_template': '{prompt}\n<START>\n', 'bot_template': '{bot_name}: {message}\n', 'user_template': '{user_name}: {message}\n', 'response_template': '{bot_name}:', 'truncate_by_message': False}
timestamp: 2024-05-06T16:38:03+00:00
model_name: lumimaid
model_eval_status: success
double_thumbs_up: 363
thumbs_up: 592
thumbs_down: 331
num_battles: 32800
num_wins: 15135
celo_rating: 1155.07
entertaining: 6.92
stay_in_character: 8.48
user_preference: 7.32
safety_score: 0.9
submission_type: basic
model_architecture: LlamaForCausalLM
model_num_parameters: 8030261248.0
best_of: 4
max_input_tokens: 512
max_output_tokens: 64
display_name: lumimaid
double_thumbs_up_ratio: 0.2822706065318818
feedback_count: 1286
ineligible_reason: None
language_model: neversleep/llama-3-lumimaid-8b
model_score: 7.573333333333333
model_size: 8B
reward_model: ChaiML/reward_gpt2_medium_preference_24m_e2
single_thumbs_up_ratio: 0.4603421461897356
thumbs_down_ratio: 0.2573872472783826
thumbs_up_ratio: 0.7426127527216174
us_pacific_date: 2024-05-06
win_ratio: 0.4614329268292683
Resubmit model
Running pipeline stage MKMLizer
Starting job with name neversleep-llama-3-lumim-8666-v2-mkmlizer
Waiting for job on neversleep-llama-3-lumim-8666-v2-mkmlizer to finish
neversleep-llama-3-lumim-8666-v2-mkmlizer: ╔═════════════════════════════════════════════════════════════════════╗
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ _____ __ __ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ / _/ /_ ___ __/ / ___ ___ / / ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ / _/ / // / |/|/ / _ \/ -_) -_) / ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ /_//_/\_, /|__,__/_//_/\__/\__/_/ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ /___/ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ Version: 0.8.10 ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ Copyright 2023 MK ONE TECHNOLOGIES Inc. ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ The license key for the current software has been verified as ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ belonging to: ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ Chai Research Corp. ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ Account ID: 7997a29f-0ceb-4cc7-9adf-840c57b4ae6f ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ Expiration: 2024-07-15 23:59:59 ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ║ ║
neversleep-llama-3-lumim-8666-v2-mkmlizer: ╚═════════════════════════════════════════════════════════════════════╝
neversleep-llama-3-lumim-8666-v2-mkmlizer: /opt/conda/lib/python3.10/site-packages/huggingface_hub/utils/_deprecation.py:131: FutureWarning: 'list_files_info' (from 'huggingface_hub.hf_api') is deprecated and will be removed from version '0.23'. Use `list_repo_tree` and `get_paths_info` instead.
neversleep-llama-3-lumim-8666-v2-mkmlizer: warnings.warn(warning_message, FutureWarning)
neversleep-llama-3-lumim-8666-v2-mkmlizer: Downloaded to shared memory in 15.800s
neversleep-llama-3-lumim-8666-v2-mkmlizer: quantizing model to /dev/shm/model_cache
neversleep-llama-3-lumim-8666-v2-mkmlizer: Saving flywheel model at /dev/shm/model_cache
neversleep-llama-3-lumim-8666-v2-mkmlizer: Loading 0: 0%| | 0/291 [00:00<?, ?it/s] Loading 0: 1%| | 2/291 [00:04<09:57, 2.07s/it] Loading 0: 57%|█████▋ | 166/291 [00:05<00:03, 37.54it/s] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
neversleep-llama-3-lumim-8666-v2-mkmlizer: quantized model in 18.773s
neversleep-llama-3-lumim-8666-v2-mkmlizer: Processed model neversleep/llama-3-lumimaid-8b in 35.564s
neversleep-llama-3-lumim-8666-v2-mkmlizer: creating bucket guanaco-mkml-models
neversleep-llama-3-lumim-8666-v2-mkmlizer: Bucket 's3://guanaco-mkml-models/' created
neversleep-llama-3-lumim-8666-v2-mkmlizer: uploading /dev/shm/model_cache to s3://guanaco-mkml-models/neversleep-llama-3-lumim-8666-v2
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /dev/shm/model_cache/flywheel_model.0.safetensors s3://guanaco-mkml-models/neversleep-llama-3-lumim-8666-v2/flywheel_model.0.safetensors
neversleep-llama-3-lumim-8666-v2-mkmlizer: loading reward model from ChaiML/reward_gpt2_medium_preference_24m_e2
neversleep-llama-3-lumim-8666-v2-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:913: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-llama-3-lumim-8666-v2-mkmlizer: warnings.warn(
neversleep-llama-3-lumim-8666-v2-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/tokenization_auto.py:757: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-llama-3-lumim-8666-v2-mkmlizer: warnings.warn(
neversleep-llama-3-lumim-8666-v2-mkmlizer: /opt/conda/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:468: FutureWarning: The `use_auth_token` argument is deprecated and will be removed in v5 of Transformers. Please use `token` instead.
neversleep-llama-3-lumim-8666-v2-mkmlizer: warnings.warn(
neversleep-llama-3-lumim-8666-v2-mkmlizer: /opt/conda/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly. To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()
neversleep-llama-3-lumim-8666-v2-mkmlizer: return self.fget.__get__(instance, owner)()
neversleep-llama-3-lumim-8666-v2-mkmlizer: Saving model to /tmp/reward_cache/reward.tensors
neversleep-llama-3-lumim-8666-v2-mkmlizer: Saving duration: 0.293s
neversleep-llama-3-lumim-8666-v2-mkmlizer: Processed model ChaiML/reward_gpt2_medium_preference_24m_e2 in 7.370s
neversleep-llama-3-lumim-8666-v2-mkmlizer: creating bucket guanaco-reward-models
neversleep-llama-3-lumim-8666-v2-mkmlizer: Bucket 's3://guanaco-reward-models/' created
neversleep-llama-3-lumim-8666-v2-mkmlizer: uploading /tmp/reward_cache to s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/config.json s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/config.json
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/tokenizer_config.json s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/tokenizer_config.json
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/special_tokens_map.json s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/special_tokens_map.json
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/merges.txt s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/merges.txt
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/vocab.json s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/vocab.json
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/tokenizer.json s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/tokenizer.json
neversleep-llama-3-lumim-8666-v2-mkmlizer: cp /tmp/reward_cache/reward.tensors s3://guanaco-reward-models/neversleep-llama-3-lumim-8666-v2_reward/reward.tensors
Job neversleep-llama-3-lumim-8666-v2-mkmlizer completed after 64.33s with status: succeeded
Stopping job with name neversleep-llama-3-lumim-8666-v2-mkmlizer
Pipeline stage MKMLizer completed in 69.12s
Running pipeline stage MKMLKubeTemplater
Pipeline stage MKMLKubeTemplater completed in 0.09s
Running pipeline stage ISVCDeployer
Creating inference service neversleep-llama-3-lumim-8666-v2
Waiting for inference service neversleep-llama-3-lumim-8666-v2 to be ready
Inference service neversleep-llama-3-lumim-8666-v2 ready after 40.291786670684814s
Pipeline stage ISVCDeployer completed in 47.84s
Running pipeline stage StressChecker
Received healthy response to inference request in 2.123077630996704s
Received healthy response to inference request in 1.133194923400879s
Received healthy response to inference request in 1.0883679389953613s
Received healthy response to inference request in 1.0845739841461182s
Received healthy response to inference request in 0.9431116580963135s
5 requests
0 failed requests
5th percentile: 0.9714041233062745
10th percentile: 0.9996965885162353
20th percentile: 1.0562815189361572
30th percentile: 1.0853327751159667
40th percentile: 1.086850357055664
50th percentile: 1.0883679389953613
60th percentile: 1.1062987327575684
70th percentile: 1.1242295265197755
80th percentile: 1.3311714649200441
90th percentile: 1.7271245479583741
95th percentile: 1.9251010894775389
99th percentile: 2.083482322692871
mean time: 1.274465227127075
Pipeline stage StressChecker completed in 6.99s
Running pipeline stage DaemonicModelEvalScorer
Pipeline stage DaemonicModelEvalScorer completed in 0.03s
Running pipeline stage DaemonicSafetyScorer
Running M-Eval for topic stay_in_character
Pipeline stage DaemonicSafetyScorer completed in 0.04s
M-Eval Dataset for topic stay_in_character is loaded
neversleep-llama-3-lumim_8666_v2 status is now deployed due to DeploymentManager action
neversleep-llama-3-lumim_8666_v2 status is now inactive due to auto deactivation removed underperforming models

Usage Metrics

Latency Metrics